Current Landscape of Late-Phase Clinical Trials for Alzheimer's Disease: Comparing Regional Variation Between Subjects in Japan and North America

Introduction Over the last few decades, numerous late-phase multi-regional clinical trials have been conducted to develop a novel treatment for Alzheimer's disease (AD), with no effective results. Objective To inform the design and interpretation of future clinical trials, the aim of this study was first to examine the current landscape of late-phase clinical trials to determine key study design characteristics, and then assess the regional variation between Japan and North America for the most utilized clinical efficacy endpoint in the most targeted stage of the disease. Methods The study design and the mechanism of action of the interventional drugs tested in the late-phase clinical trials initiated in the last 5 years (2014-2018) were assessed based on the records in ClinicalTrials.gov database. The regional variation of the most utilized clinical efficacy endpoint in the most targeted population was assessed using data from two similarly designed observational studies conducted in Japan (Japanese Alzheimer's Disease Neuroimaging Initiative, J-ADNI) and North America (Alzheimer's Disease Neuroimaging Initiative, ADNI). For the most utilized clinical efficacy endpoint, the change from baseline (CFB) at Month 6, Year 1 and Year 2 was estimated using the growth curve model with a random intercept and slope, including gender as a fixed factor and age, apolipoprotein E epsilon 4 genotype and years of education as covariates. Results Of 48 Phase III trials that were initiated during the study period, 25 were disease-modifying treatment trials in which individuals with early AD were the most studied (56%) and Clinical Dementia Rating-Sum of Boxes (CDR-SB) was the most frequently utilized primary clinical efficacy endpoint (64%). The baseline characteristics of the early AD population between J-ADNI and ADNI were generally comparable, except for years of education. When comparing CDR-SB in early AD, J-ADNI had generally better baseline scores and the overall progression was similar (CFB at Year 2, ADNI 2.7 and J-ADNI 2.3, p = 0.190), despite slower progression in functional domains (CFB at Year 2, ADNI 1.4 and J-ADNI 1.0, p = 0.031). Conclusion Over the years, the target population has shifted toward early stage of the disease, wherein the clinical progression is slower and difficult to measure. Moreover, our results suggest that regional variation could have an impact on functional measurements due to cultural differences in pivotal clinical trials. Therefore, caution should be exercised according to the characteristics of the endpoint used.